It sounds similar to the more thoughtful critiques of AI in programming. Treating AI as a lines of code (LOC) generator, where actual human beings end up not en

It sounds similar to the more thoughtful critiques of AI in programming. Treating AI as a lines of code (LOC) generator, where actual human beings end up not engaging with the domain the software is working in, and there is then little to no understanding or care to understand the resulting code, we call "slop." The term slop is definitely pejorative, but since most software is commercial, it ends up raising questions as to how much quality and domain knowledge matter if the goal is primarily marketplace success. Then we get those thorny questions of what is software for, anyway -- do we really accept the premise that "marketplace success" is the best, and perhaps only relevant, criteria? The declaration here, as I understand it, is taking the premise that the point of mathematics _is_ the domain knowledge, and that using "solves math problems" as a benchmark is exactly the kind of shallow, domain-eroding approach to AI that in software we label "slop."